Background: Depression has been correlated with suboptimal adherence to antidiabetic drugs (ADs). Most studies on this topic were cross-sectional; thus, the directionality of this relationship could not be established. The objective of this study was to measure the association between incident depression and AD nonadherence among newly treated patients with diabetes.
Methods: We performed a population-based cohort study among new AD users using the Quebec public health insurance data. To avoid immortal time bias, we carried out depression diagnosis-time distribution matching by assigning a date of depression diagnosis to individuals without depression. Nonadherence (i.e.,<90% of days covered by≥1 AD) during the year following depression diagnosis (real or assigned date) was the outcome. Multivariate logistic regression analyses that adjusted for baseline adherence and other confounders were used to estimate the adjusted effect of depression on AD nonadherence.
Results: Between 2000 and 2006, we identified 3,106 new AD users with a subsequent diagnosis of depression and 70,633 without depression, of which 52% and 49% became non-adherent to AD treatment, respectively. Among patients with depression, 52.0% were considered AD non-adherent in the year after depression diagnosis compared with 49.0% of matched patients without depression. Depression was associated with AD nonadherence after accounting for baseline adherence and other confounders with an adjusted odds ratio of 1.24 (95% confidence interval: 1.13-1.37).
Conclusions: The results suggest that depression is an independent risk factor for AD nonadherence. Patients with type 2 diabetes and depression might benefit from adherence-enhancing interventions.
Keywords: Depression; Healthcare administrative claims; Immortal time bias; Medication adherence; Type 2 diabetes mellitus.
Copyright © 2017 Elsevier Masson SAS. All rights reserved.